ECON 323 Project

[Shared anonymously by request. Two student names redacted]

We found the data needed, cleaned the data, merged the data, and brainstormed graphs and plots that might be used for the data visualization together. Then, each person wrote some different parts of the code with the explanation for markdown parts. Afterwards, we explained to each other the work we've done and why we code like this to create such kind of visualization, and finally we unified each part to complete the entire project.

The main objective of this analysis is to investigate the possible correlation between a country's Olympic medal count and its economic performance. To be precise, the research question being addressed is:

Whether there exists a relationship between the number of Olympic medals won by a country and its GDP per capita, as well as the underlying factors that may contribute to this association.

Background:¶

The Olympic Games are a global multi-sport event that brings together athletes from around the world to compete against each other. The first modern Olympics took place in Athens, Greece in 1896 and have since grown into a phenomenon that captures the attention of millions of people worldwide.

The Games include both summer and winter sports with a variety of disciplines, from track and field to gymnastics, swimming, skiing, and more. The Games are a symbol of unity, excellence, and perseverance, inspiring athletes and spectators alike to pursue their dreams and reach their full potential. Athletes come from all walks of life to represent their countries and compete for medals. When we look at the public medal tally of previous Olympic Games, we can see that most of the top-ranking countries are developed nations. This raises the question of whether a country's economic performance can also affect its performance in sporting events on such a global stage as the Olympic Games, which unite nations, promote diplomacy, and showcase national strength. Given the tremendous influence of the Games, we believe that examining the relationship between economic performance and Olympic success can provide valuable insights into the interplay between sports and the economy, and the ways in which countries leverage their resources to achieve sporting excellence.

Data selection:¶

For the indicator of countries’ economic performance, we think Gross Domestic Product could be a good consideration. GDP is a measure of the total value of all goods and services produced within a country's borders in a given period, typically a year. GDP per capita is the GDP divided by the total population of a country. It is a measure of the average economic output per person and is often used to compare living standards between countries. GDP per capita provides a more accurate picture of the standard of living in a country, as it takes into account population size. In this research, we will use GDP per capita, instead of using data on GDP, to predict earnings and show the standard of living across countries. While GDP per capita alone still does not provide a complete picture of a country's economic performance, we will use it as an accurate indicator in this study, ignoring other factors such as income inequality, access to healthcare and education, and social welfare programs.

To ensure accurate results, we collected data from three Summer Olympic Games (2008 Beijing, 2012 London, and 2016 Rio) and three Winter Olympic Games (2010 Vancouver, 2014 Sochi, and 2018 Pyeongchang). However, we made a conscious decision to exclude the data from the two most recent Olympic Games (2020 Tokyo and 2022 Beijing) due to the unprecedented COVID-19 pandemic that affected the global economy.

The outbreak of COVID-19 in early 2020 led to significant changes in economic trends worldwide. The pandemic resulted in lockdowns and travel restrictions, which severely impacted international trade, supply chains, and consumer demand. Countries that rely heavily on tourism, such as Japan, were hit particularly hard due to the postponement of the 2020 Tokyo Olympics. Moreover, many countries faced a decline in their GDP as a result of reduced economic activities and decreased productivity.

Therefore, to ensure that our study results reflect the reality of countries' economic performances, we decided to exclude the data from the recent two Olympic Games. By doing so, we aim to eliminate the potential distortion caused by the COVID-19 pandemic and ensure that the outcomes are closer to the realized fact. Although excluding the data from the two most recent Olympic Games may limit the scope of our analysis, we believe that it will provide more accurate and meaningful results.

Due to time constraints, we were unable to obtain precise data on the number of ski resorts in countries and ski population share in European countries in a time-series manner for our analysis of the underlying factors contributing to the relationship between the number of Olympic medals won by a country and its GDP per capita. Obtaining specific values for these two types of data for the range of Olympic years selected for our study (2008-2018) proved to be extremely difficult, then we decided to assume that the two data points did not change significantly over time and used cross-sectional data to analyze the relevant potential factors. (from just a few of these time points we found, the two types of data really do not change much). Anyway, we will continue to enhance the flavor/richness of this project in the very near future.

Note: in the following analysis, we refer to high GDP per capita countries as high-income countries.

Methodology:¶

  1. General visualizations to show the economic performance and Olympic success across countries (including GDP per capita, medal amounts, growth rate, and medal ranking by countries, etc.)
  2. Visualization of the trend between the number of Olympic medals and countries’ GDP (for the top 15 countries in the medal standings of the six selected Olympic Games) - Presentation of estimation results from regressions
  3. Exploring the underlying reasons that might explain that correlation:
  • Economic perspective
  • Participant demographic perspective
  • Cultural perspective
  • Geographical perspective
  1. Conclusion, discussion of the limitations of our study, and areas for further research.

The following visualizations have some interaction, for example, the scatter chart where you can hover over a point, and specific information about that point will be displayed. So, remember to move your mouse around the chart to explore it!

GDP per capita Dataset
Presentation of GDP per capita by country from 2008 to 2018

In [1]:
import numpy as np # linear algebra
import pandas as pd # data processing, CSV file I/O (e.g. pd.read_csv)
import seaborn as sns
import matplotlib.pyplot as plt
In [2]:
df = pd.read_excel('323_data.xlsx', sheet_name='GDP per capita')
df.head(11).set_index('year')
Out[2]:
United States of America People's Republic of China Russian Federation Great Britain Australia France Germany Republic of Korea Cuba Italy ... Netherlands Hungary Azerbaijan Norway Austria Sweden Switzerland Czech Republic Poland Slovenia
year
2008 55271.65625 7412.874512 24887.85352 43767.69531 44743.66016 42996.21094 47643.22266 32275.12500 NaN 44623.60156 ... 53848.25391 25835.54883 12672.3760 63570.33984 53166.05469 47963.4922 66598.8594 34580.4414 22444.90820 36075.4688
2009 53399.36719 8069.354492 22939.69336 41592.92969 44649.16016 41546.49219 45044.48828 32363.96875 NaN 42074.92188 ... 51607.60547 24168.98438 13565.8555 61689.55859 51030.72656 45492.7695 64404.0781 32782.6133 23064.95313 33052.2266
2010 54315.91406 8884.587891 23961.22070 42147.76563 44914.98047 42147.67188 46999.23828 34394.49219 NaN 42664.35547 ... 52032.98828 24486.04102 14082.2383 61353.47656 51843.42969 47791.3711 65819.7344 33483.1367 23996.14063 33350.6133
2011 54758.83203 9680.097656 24972.07813 42429.21094 45389.17188 42864.01953 49757.92578 35388.97656 NaN 42892.30469 ... 52594.23047 25012.29492 13913.8369 61157.12891 53179.14844 48947.4375 66342.0078 34002.1914 25124.27539 33568.0664
2012 55581.41406 10370.726560 25933.29297 42754.53906 46350.92188 42790.63281 49872.44922 36049.19531 NaN 41501.71094 ... 51860.05469 24825.58008 14027.4688 61990.69141 53297.44531 48300.8438 66440.2891 33688.1016 25457.20703 32613.4902
2013 56214.42969 11101.938480 26332.39648 43271.82813 46744.63281 42816.26953 49954.17188 37021.13281 NaN 40268.11328 ... 51640.07813 25343.71289 14651.6943 61878.97656 52997.75391 48462.1602 66877.1641 33661.4688 25759.33984 32234.0254
2014 57213.26953 11851.404300 26057.15625 44239.15234 47232.87500 43021.39453 50845.52734 37967.47656 NaN 39898.52734 ... 52186.99609 26485.41211 14867.9434 62390.12891 52932.89844 49258.9961 67682.6875 34386.7031 26649.57813 33093.7500
2015 58540.29688 12612.351560 25488.09570 45041.05859 47569.29297 43345.78516 51159.29688 38828.74219 NaN 40247.82813 ... 52974.11719 27531.64453 14852.6113 62987.51563 52873.85938 50928.9648 68025.9219 36168.4180 27797.05859 33799.7266
2016 59112.24219 13399.137700 25490.70898 45712.78516 48116.03125 43705.14844 51879.67188 39814.66016 NaN 40837.73828 ... 53847.82813 28218.27930 14232.1924 63104.08203 53345.74219 51334.4336 68663.2422 37014.9609 28682.69336 34853.0430
2017 60109.65625 14243.532230 25926.44336 46372.38672 48398.55078 44577.06641 53071.45703 40957.41797 NaN 41581.12109 ... 55088.63281 29501.12109 14121.4072 64050.75781 54172.98828 51947.9531 69103.5703 38824.8867 30064.50391 36507.5547
2018 61585.75781 15133.996090 26656.41016 46853.19922 49029.10547 45284.00391 53486.83984 41965.89063 NaN 42045.92188 ... 56060.91406 31121.66602 14209.5928 64341.25781 55258.37109 52349.2930 70597.2891 39932.9922 31674.13086 37984.1328

11 rows × 25 columns

Resource: https://ourworldindata.org/grapher/gdp-per-capita-worldbank?tab=chart&yScale=log&time=latest&country=

In [3]:
import pandas as pd
import plotly.express as px
import plotly.offline as py
import numpy as np


def generate_hover_data(row):
    hover_data = {}
    if row['Year'] in [2008, 2012, 2016, 2010, 2014, 2018]:
        for part in ['Value', 'Gold', 'Silver', 'Bronze', 'Total']:
            if row[part] != '':
                hover_data[part] = row[part]
            else:
                hover_data[part] = '0'
    else:
        # if row['Value'] != '':
        #     hover_data['Value'] = row['Value']
        # else:
        #     hover_data['Value'] = '0'
        for part in ['Value', 'Gold', 'Silver', 'Bronze', 'Total']:
            if row[part] != '':
                hover_data[part] = row[part]
            else:
                hover_data[part] = '0'
    hover_data['0'] = 1
    return hover_data

excel_pth = '323_data.xlsx'
medal_amount_series = pd.read_excel(excel_pth, sheet_name=['figure1'], header=1)
medal_amount_series = medal_amount_series['figure1']
GDP_series = pd.read_excel(excel_pth, sheet_name=['GDP per capita'], index_col=0, header=0)
GDP_series = GDP_series['GDP per capita']
a = medal_amount_series.iloc[:-2]
a = a.groupby(['Olympic name', 'Country'])['Gold', 'Silver', 'Bronze', 'Total'].max()
b = GDP_series
# b = b.iloc[:-2]
b = b.replace(np.nan, "NaN")
b = b.stack().reset_index()
b.columns = ['Year', 'Country', 'Value']
b['Gold'] = None
b['Silver'] = None
b['Bronze'] = None
b['Total'] = None
bb = b.groupby(['Year', 'Country']).max()
# print(bb)
year_country = {}
for info in bb.index:
    (name_year, country) = info
    if name_year not in year_country.keys():
        year_country[name_year] = []
    year_country[name_year].append(country)
for info in a.index:
    (name, country) = info
    name_year = int(name[:4])
    # print(year_country)
    if name_year == 2018 and country == 'ROC':
        ccc = 0
    if country not in year_country[name_year]:
        for part in ['Value', 'Gold', 'Silver', 'Bronze', 'Total']:
            bb.loc[(name_year, country), part] = None
    for part in ['Gold', 'Silver', 'Bronze', 'Total']:
        bb[part][(name_year, country)] = a[part][info]
bb = bb.reset_index()
# bb['Year'] = pd.to_datetime(bb['Year'].astype(str), format='%Y').dt.year
bb['Year'] = pd.to_datetime(bb['Year'].astype(float), format='%Y').dt.year
bb['Value'] = bb['Value'].astype(float)
max_color_value = int(max(bb['Value']))+2
bb['Size'] = bb['Value']
bb['Size'] = bb['Size'].fillna(0)
sizemax = max(bb['Size'])
bb['Size'] = (bb['Size']/1000)**1.5

bb['Value'] = bb['Value'].fillna('null')
bb['Gold'] = bb['Gold'].fillna('null')
bb['Silver'] = bb['Silver'].fillna('null')
bb['Bronze'] = bb['Bronze'].fillna('null')
bb['Total'] = bb['Total'].fillna('null')

# hover_data=bb.apply(generate_hover_data, axis=1)

# fig = px.scatter_geo(bb, locations="Country", locationmode='country names',
#                      color_discrete_sequence=['blue'], size_max=30, hover_name="Country", hover_data=hover_data,
#                      range_color=[0, max_color_value],
#                      projection="natural earth", animation_frame='Year')

hover_data = [
    ['Value', 'Gold', 'Silver', 'Bronze', 'Total'],
    ['Value'],
] * (278 // 2)
# fig = px.scatter_geo(bb, locations="Country", locationmode='country names',
#                      color_discrete_sequence=['blue'], size_max=30, hover_name="Country", hover_data=['Value', 'Gold', 'Silver', 'Bronze', 'Total'],
#                      range_color=[0, max_color_value],
#                      projection="natural earth", animation_frame='Year')
fig = px.scatter_geo(bb, locations="Country", locationmode='country names',
                     color_discrete_sequence=['blue'], size='Size', hover_name="Country", hover_data=['Value', 'Gold', 'Silver', 'Bronze', 'Total'],
                     range_color=[0, max_color_value],
                     projection="natural earth", animation_frame='Year',
                     width=1000,
                     height=800)

test_list = []
for i in range(len(bb)):
    year = bb.loc[i, 'Year']
    if year in [2008, 2012, 2016, 2010, 2014, 2018]:
        text = f"Gold: {bb.loc[i, 'Gold']}, Silver: {bb.loc[i, 'Silver']}, Bronze: {bb.loc[i, 'Bronze']}, Total: {bb.loc[i, 'Total']}, Value: {bb.loc[i, 'Value']}"
    else:
        text = f"Value: {bb.loc[i, 'Value']}"
    test_list.append(text)
fig.data[0].text = test_list
fig.update_traces(text=fig.data[0].text, textposition='top center', hovertemplate='%{text}<br>')
fig.update(layout_coloraxis_showscale=False,layout_title={
        'text': "GDP per capita and Olympic medals by country from 2008 to 2018",'x': 0.5,'y': 0.95})
py.offline.iplot(fig)

/tmp/ipykernel_247/331117175.py:34: FutureWarning:

Indexing with multiple keys (implicitly converted to a tuple of keys) will be deprecated, use a list instead.

/tmp/ipykernel_247/331117175.py:62: SettingWithCopyWarning:


A value is trying to be set on a copy of a slice from a DataFrame

See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy

/tmp/ipykernel_247/331117175.py:62: SettingWithCopyWarning:


A value is trying to be set on a copy of a slice from a DataFrame

See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy

/tmp/ipykernel_247/331117175.py:62: SettingWithCopyWarning:


A value is trying to be set on a copy of a slice from a DataFrame

See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy

/tmp/ipykernel_247/331117175.py:62: SettingWithCopyWarning:


A value is trying to be set on a copy of a slice from a DataFrame

See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy


                                                   Figure 1

With this world map, we can visualize the geographical distribution of the top 15 winning countries of the each six selected Olympic Games(2008-2018), the total number of medals of each Olympics, and the GDP per capita of each countries during this decade of years 2008-2018. A quick glance at the map reveals that gold medalists are spread across the globe, with a higher concentration in Europe, North America, and East Asia, primarily in affluent nations. It might make us start to wonder if there is any relationship between sports performance and national income and development.

Notice: Each dot shown on this world map represents a country. When it is a non-Olympic year, for example 2009, the number of medals is displayed null. The size of the dot indicates the corresponding magnitude of GDP per capita for each country. The "value" in the information box represents gdp per capita, but if this value is used as the size of the dot directly, we found that there is a lot of overlap and the size distinction is not obvious, so we added a new variable-"size", which is transformed by "value", to be the standard of size representation of dots.

In addition, because the top 15 countries are mostly European countries, the images of the European region will be more overlapping, so please zoom in for better results.

Medal amount Dataset

Example: Presentation of Medal amount of The 2008 Summer Olympics are shown below.

In [4]:
df = pd.read_excel('323_data.xlsx', sheet_name='figure2',header = 1)
df.head(10).set_index('Rank')

/opt/conda/lib/python3.10/site-packages/pandas/core/indexes/base.py:6982: FutureWarning:

In a future version, the Index constructor will not infer numeric dtypes when passed object-dtype sequences (matching Series behavior)
Out[4]:
Country Gold Silver Bronze Total
Rank
1 USA 36 39 37 112
2 China 48 22 30 100
3 Russian 24 13 23 60
4 Great Britain 19 13 19 51
5 Australia 14 15 17 46
6 France 7 16 20 43
7 Germany 16 11 14 41
8 Korea 13 11 8 32
9 Cuba 3 10 17 30
10 Italy 8 9 10 27
In [5]:
summer2008 = df.iloc[0:15, :]
summer2012 = df.iloc[18:33, :]
summer2016 = df.iloc[36:51, :]
winter2010 = df.iloc[54:69, :]
winter2014 = df.iloc[72:87, :]
winter2018 = df.iloc[90:105, :]
In [6]:
import pandas as pd
summer08 = pd.DataFrame({'Country': summer2008.iloc[:, 1],'Gold': summer2008.iloc[:, 2],'Silver': summer2008.iloc[:, 3],
                         'Bronze': summer2008.iloc[:, 4],'Total': summer2008.iloc[:, 5]}).set_index('Country')
summer12 = pd.DataFrame({'Country': summer2012.iloc[:, 1],'Gold': summer2012.iloc[:, 2],'Silver': summer2012.iloc[:, 3],
                         'Bronze': summer2012.iloc[:, 4],'Total': summer2012.iloc[:, 5]}).set_index('Country')
summer16 = pd.DataFrame({'Country': summer2016.iloc[:, 1],'Gold': summer2016.iloc[:, 2],'Silver': summer2016.iloc[:, 3],
                         'Bronze': summer2016.iloc[:, 4],'Total': summer2016.iloc[:, 5]}).set_index('Country')



colors = [ (0.804, 0.498, 0.196),(0.75, 0.75, 0.75), (0.855, 0.647, 0.125),(1.0, 0.0, 0.0)]

fig, axes = plt.subplots(nrows=1, ncols=3, figsize=(15, 7))
# plt.subplots_adjust(wspace=0.5)
summer08.sort_values(by="Total", ascending=True).tail(15).plot(kind="barh", ax=axes[0], color=colors)
summer12.sort_values(by="Total", ascending=True).tail(15).plot(kind="barh", ax=axes[1], color=colors)
summer16.sort_values(by="Total", ascending=True).tail(15).plot(kind="barh", ax=axes[2], color=colors)

# Set background color
# axes[0].set_facecolor('white')

# Add grid
axes[0].grid(True, axis='x', linestyle='--', color='gray', linewidth=0.5)
axes[0].spines['right'].set_visible(False)
axes[0].spines['top'].set_visible(False)
axes[0].set_title("Medal of The 2008 Summer Olympics",fontsize=14)
axes[1].grid(True, axis='x', linestyle='--', color='gray', linewidth=0.5)
axes[1].spines['right'].set_visible(False)
axes[1].spines['top'].set_visible(False)
axes[1].set_title("Medal of The 2012 Summer Olympics",fontsize=14)
axes[2].grid(True, axis='x', linestyle='--', color='gray', linewidth=0.5)
axes[2].spines['right'].set_visible(False)
axes[2].spines['top'].set_visible(False)
axes[2].set_title("Medal of The 2016 Summer Olympics",fontsize=14)
plt.tight_layout()

                                              Figure 2 (1)

From 2008, 2012, and 2016 three Summer Olympics medal count bar charts, which are shown above, we can recognize that the United States, China, Britain, and Russia, which have occupied the top fourth in three consecutive games, were really dominating the Olympic medal tally. In specific, China, the United States have been located in either the first, or second-place positions. For a country like China, which does not have a high GDP per capita, how could it perform well in the Olympic Games?

In [7]:
winter10 = pd.DataFrame({'Country': winter2010.iloc[:, 1],'Gold': winter2010.iloc[:, 2],'Silver': winter2010.iloc[:, 3],
                         'Bronze': winter2010.iloc[:, 4],'Total': winter2010.iloc[:, 5]}).set_index('Country')
winter14 = pd.DataFrame({'Country': winter2014.iloc[:, 1],'Gold': winter2014.iloc[:, 2],'Silver': winter2014.iloc[:, 3],
                         'Bronze': winter2014.iloc[:, 4],'Total': winter2014.iloc[:, 5]}).set_index('Country')
winter18 = pd.DataFrame({'Country': winter2018.iloc[:, 1],'Gold': winter2018.iloc[:, 2],'Silver': winter2018.iloc[:, 3],
                         'Bronze': winter2018.iloc[:, 4],'Total': winter2018.iloc[:, 5]}).set_index('Country')

colors = [(0.902, 0.902, 0.997), (0.695, 0.695, 0.993), (0.488, 0.488, 0.989),(0.282, 0.282, 0.985), (0.078, 0.078, 0.980)]
fig, axes = plt.subplots(nrows=1, ncols=3, figsize=(15, 7))

plt.subplots_adjust(wspace=0.6)
winter10.sort_values(by="Total", ascending=True).tail(15).plot(kind="barh", ax=axes[0], color=colors)
winter14.sort_values(by="Total", ascending=True).tail(15).plot(kind="barh", ax=axes[1], color=colors)
winter18.sort_values(by="Total", ascending=True).tail(15).plot(kind="barh", ax=axes[2], color=colors)
# Add grid
axes[0].grid(True, axis='x', linestyle='--', color='gray', linewidth=0.5)
axes[0].spines['right'].set_visible(False)
axes[0].spines['top'].set_visible(False)
axes[0].set_title("Medal of The 2010 Winter Olympics",fontsize=14)
axes[1].grid(True, axis='x', linestyle='--', color='gray', linewidth=0.5)
axes[1].spines['right'].set_visible(False)
axes[1].spines['top'].set_visible(False)
axes[1].set_title("Medal of The 2014 Winter Olympics",fontsize=14)
axes[2].grid(True, axis='x', linestyle='--', color='gray', linewidth=0.5)
axes[2].spines['right'].set_visible(False)
axes[2].spines['top'].set_visible(False)
axes[2].set_title("Medal of The 2018 Winter Olympics",fontsize=14)
plt.tight_layout()

                                               Figure 2 (2)

Resource: https://olympics.com/en/olympic-games

In contrast to the Summer Olympics, the position at the top of the medal standings at the Winter Olympics seems to be more volatile. China no longer occupies the top two positions, and its ranking has dropped. On the other hand, Norway and Canada, etc have shown relatively better performances at the Winter Olympics, and have rushed to the top of the medal standings. Overall, the Winter Olympics have showcased a higher degree of variability in medal standings compared with Summer Olympics, with different countries performing well in different Olympic periods.

Line chart of ranking for countries in three summer Olympics and three winter Olympics(Only countries ranked in the top 15 in all three Olympic Games are included)

In [11]:
df = pd.read_excel('323_data.xlsx', sheet_name='figure3')
df.head(11).set_index('Country')
Out[11]:
2008 2012 2016 Country 2010 2014 2018
Country
USA 1 1 1 | | USA 1 2 3
China 2 2 2 | | Germany 2 6 2
Russian 3 3 4 | | Canada 3 4 3
Great Britain 4 4 3 | | Norway 4 3 1
Australia 5 8 8 | | Austria 5 7 11
France 6 7 6 | | Korea 7 13 6
Germany 7 5 5 | | China 8 11 14
Korea 8 9 11 | | Sweden 8 9 10
Italy 10 10 9 | | France 10 8 9
Japan 11 6 7 | | Switzerland 11 10 8
Canada 13 15 10 | | Netherlands 12 5 5
In [12]:
import pandas as pd
import matplotlib.pyplot as plt

# Data for summer Olympics 
summer_data = {'Country': ['USA', 'China', 'Russian', 'Great Britain', 'Australia', 'France', 'Germany', 'Korea', 'Italy', 'Japan', 'Canada'],
               2008: [1, 2, 3, 4, 5, 6, 7, 8, 10, 11, 13],
               2012: [1, 2, 3, 4, 8, 7, 5, 9, 10, 6, 15],
               2016: [1, 2, 4, 3, 8, 6, 5, 11, 9, 7, 10]}

# Data for winter Olympics 
winter_data = {'Country': ['USA', 'Germany', 'Canada', 'Norway', 'Austria', 'Korea', 'China', 'Sweden', 'France', 'Switzerland', 'Netherlands', 'Czech Republic'],
               2010: [1, 2, 3, 4, 5, 7, 8, 8, 10, 11, 12, 13],
               2014: [2, 6, 4, 3, 7, 13, 11, 9, 8, 10, 5, 12],
               2018: [3, 2, 3, 1, 11, 6, 14, 10, 9, 8, 5, 15]}

# To convert data into DataFrame format
summer_df = pd.DataFrame(summer_data)
winter_df = pd.DataFrame(winter_data)

# To creat subplot
fig, (ax1, ax2) = plt.subplots(nrows=1, ncols=2, figsize=(16, 8))

# Creating line chart of ranking for countries in summer Olympics 
for i, country in enumerate(summer_df['Country']):
    if country == 'China':
        ax1.plot(summer_df.columns[1:], summer_df.iloc[i, 1:], color='red', linewidth=2, label=country)
    else:
        ax1.plot(summer_df.columns[1:], summer_df.iloc[i, 1:], linewidth=2, label=country)
ax1.invert_yaxis()

ax1.legend(loc='lower right')
ax1.set_title('Summer Olympics Country Ranking over Time')


ax1.set_facecolor('#FFFACD')

# Creating line chart of ranking for countries in winter Olympics 
for i, country in enumerate(winter_df['Country']):
    if country == 'China':
        ax2.plot(winter_df.columns[1:], winter_df.iloc[i, 1:], color='red', linewidth=2, label=country)
    else:
        ax2.plot(winter_df.columns[1:], winter_df.iloc[i, 1:], linewidth=2, label=country)
        
ax2.invert_yaxis()
ax2.legend(loc='lower right')
ax2.set_title('Winter Olympics Country Ranking over Time')

ax2.set_facecolor('#ADD8E6')

                                                Figure 3

Resource: https://olympics.com/en/olympic-games

As we can see from the line chart, the rankings for the Winter Olympics are indeed more volatile than those for the Summer Olympics on a country basis. Although some countries may have been in the top 15 only once in the three Olympic Games and therefore did not change their ranking, we can see that more than ten countries were in the top 15 in all three Olympic Games, thus highlighting the relative stability of top-performing countries (these countries are also high-income countries). In addition, the countries leading the medal standings in the Winter Olympics differ from those in the Summer Olympics, suggesting that a country's strengths and advantages may differ for the winter and summer Olympics.

In [13]:
df = pd.read_excel('323_data.xlsx', sheet_name='figure4',header = 0)
df.head(10)
Out[13]:
2008 Medal Amount GDP per capita
0 USA 112 55271.65625
1 China 100 7412.874512
2 Russian Federation 60 24887.85352
3 Great Britain 51 43767.69531
4 Australia 46 44743.66016
5 France 43 42996.21094
6 Germany 41 47643.22266
7 Korea 32 32275.125
8 Italy 27 44623.60156
9 Japan 25 38781.19922

Scatter plots on the number of medals versus national GDP per capita in Summer Olympics

In [16]:
s08_countries = df.iloc[:14,0];s08_medal = df.iloc[:14,1];s08_gdp = df.iloc[:14,2]
s12_countries = df.iloc[16:33,0];s12_medal = df.iloc[16:33,1];s12_gdp = df.iloc[16:33,2]
s16_countries = df.iloc[33:50,0];s16_medal = df.iloc[33:50,1];s16_gdp = df.iloc[33:50,2]
In [17]:
import plotly.graph_objs as go
from plotly.subplots import make_subplots

# subplot
fig = make_subplots(rows=1, cols=3)

# 2008
fig.add_trace(
    go.Scatter(x=s08_gdp, y=s08_medal, mode='markers', marker_color='lightcoral', hovertext=s08_countries),
    row=1, col=1)
fig.update_xaxes(title_text="GDP per capita", row=1, col=1)
fig.update_yaxes(title_text="Medal Amount", row=1, col=1)
# fig.update_layout(title_text="2008 Summer Olympics Medal Amount vs GDP per capita", showlegend=False)

# 2012
fig.add_trace(
    go.Scatter(x=s12_gdp, y=s12_medal, mode='markers', marker_color='lightcoral', hovertext=s12_countries),
    row=1, col=2)
fig.update_xaxes(title_text="GDP per capita", row=1, col=2)
fig.update_yaxes( row=1, col=2)
# fig.update_layout(title_text="2012 Summer Olympics Medal Amount vs GDP per capita", showlegend=False)
#
# 2016
fig.add_trace(
    go.Scatter(x=s16_gdp, y=s16_medal, mode='markers', marker_color='lightcoral', hovertext=s16_countries),
    row=1, col=3)
fig.update_xaxes(title_text="GDP per capita", row=1, col=3)
fig.update_yaxes(row=1, col=3)
# fig.update_layout(title_text="2008|2012|2016 Summer Olympics Medal Amount vs GDP per capita", showlegend=False,height=400,width=850)
fig.update_layout(title={
        'text': "2008|2012|2016 Summer Olympics Medal Amount vs GDP per capita",'x': 0.5,'y': 0.85},
        showlegend=False,height=400,width=850)
# plot
fig.show()

                                      Figure 4 (1)

Scatterplots on the number of medals versus national GDP per capita in Winter Olympics

In [18]:
s10_countries = df.iloc[48:64,0];s10_medal = df.iloc[48:64,1];s10_gdp = df.iloc[48:64,2]
s14_countries = df.iloc[65:81,0];s14_medal = df.iloc[65:81,1];s14_gdp = df.iloc[65:81,2]
s18_countries = df.iloc[82:98,0];s18_medal = df.iloc[82:98,1];s18_gdp = df.iloc[82:98,2]
# s18_medal
In [19]:
# subplot
fig = make_subplots(rows=1, cols=3)

# 2010
fig.add_trace(
    go.Scatter(x=s10_gdp, y=s10_medal, mode='markers', marker_color='blue', hovertext=s10_countries),
    row=1, col=1)
fig.update_xaxes(title_text="GDP per capita", row=1, col=1)
fig.update_yaxes(title_text="Medal Amount", row=1, col=1)

# 2014
fig.add_trace(
    go.Scatter(x=s14_gdp, y=s14_medal, mode='markers', marker_color='blue', hovertext=s14_countries),
    row=1, col=2)
fig.update_xaxes(title_text="GDP per capita", row=1, col=2)
fig.update_yaxes( row=1, col=2)
#
# 2018
fig.add_trace(
    go.Scatter(x=s18_gdp, y=s18_medal, mode='markers', marker_color='blue', hovertext=s18_countries),
    row=1, col=3)
fig.update_xaxes(title_text="GDP per capita", row=1, col=3)
fig.update_yaxes(row=1, col=3)
# fig.update_layout(title_text="2008|2012|2016 Summer Olympics Medal Amount vs GDP per capita", showlegend=False,height=400,width=850)
fig.update_layout(title={
        'text': "2010|2014|2018 Winter Olympics Medal Amount vs GDP per capita",'x': 0.5,'y': 0.85},
        showlegend=False,height=400,width=850)
# plot
fig.show()

                                      Figure 4 (2)

In figure4, we made scatter plots about the number of medals versus the country's GDP per capita, both from the Summer and Winter Olympics. We can see that there is a weak positive correlation between these two variables, especially from Winter Olympic data. Thus, as the country's GDP per capita increases, the number of medals won by that country also increases, with some outliners indicated.

GDP per capita growth rate dataset for countries that have been top 15 on medels standings in six Olympic Games¶
In [20]:
df = pd.read_excel('323_data.xlsx', sheet_name='GDP per capita growth rate')
df.head(11).set_index('year')

/opt/conda/lib/python3.10/site-packages/pandas/core/indexes/base.py:6982: FutureWarning:

In a future version, the Index constructor will not infer numeric dtypes when passed object-dtype sequences (matching Series behavior)
Out[20]:
United States of America People's Republic of China Russian Federation Great Britain Australia France Germany Republic of Korea Cuba Italy ... Netherlands Hungary Azerbaijan Norway Austria Sweden Switzerland Czech Republic Poland Slovenia
year
2008 -0.820368 9.090280 5.246210 -0.937078 1.513510 -0.302484 1.152030 2.233752 4.054764 -1.615941 ... 1.773357 1.180953 8.457450 -0.768153 1.143307 -1.223070 1.509951 1.838240 4.185772 3.346280
2009 -3.450016 8.855948 -7.827750 -5.230035 -0.207409 -3.370397 -5.454577 0.275269 1.392178 -5.711508 ... -4.161039 -6.453067 7.050603 -2.958585 -4.016340 -5.151263 -3.510559 -5.198982 2.762525 -8.380320
2010 1.860292 10.102833 4.453096 1.630274 0.628837 1.448245 4.339607 6.274020 2.324767 1.400915 ... 0.824260 1.304959 3.806491 -0.544794 1.592577 5.052679 2.173638 2.136880 3.229049 0.902771
2011 0.814519 8.953816 4.218723 0.280839 0.978468 1.700610 5.869636 2.891412 2.726846 0.534287 ... 1.078628 2.155897 -1.195843 -0.320026 2.576448 2.418989 0.687340 1.550190 4.985722 0.652023
2012 1.533102 7.134524 3.849152 0.745479 2.103808 -0.171161 0.230161 1.865599 2.918528 -3.242060 ... -1.395919 -0.738926 0.816684 1.362985 0.222451 -1.321002 0.110597 -0.923735 1.545464 -2.843705
2013 1.138692 7.050734 1.538974 1.140213 0.828326 0.058150 0.163871 2.696151 2.636075 -2.972404 ... -0.424179 2.083241 4.450020 -0.180216 -0.562299 0.333985 0.628361 -0.079053 0.917452 -1.163521
2014 1.540381 6.750760 -1.045262 2.442466 1.060337 0.484564 1.784342 2.556220 0.954421 -0.917814 ... 1.059101 4.513368 1.475937 0.826056 -0.122371 1.644244 1.108694 2.154500 3.914706 2.667134
2015 1.953004 6.420736 -2.183885 1.584979 0.693066 0.754024 0.617105 2.268429 4.365679 0.875477 ... 1.508266 3.954283 -0.103121 0.957501 -0.111540 3.390176 0.494130 5.181403 4.452884 2.133258
2016 0.933375 6.238215 0.010248 1.393849 1.138424 0.829057 1.408102 2.539145 0.494992 1.465690 ... 1.649316 2.503053 -4.177172 0.185060 0.892469 0.796146 0.960132 2.340552 2.997773 3.116347
2017 1.597136 6.301863 1.709386 1.749954 0.604963 1.994998 2.297206 2.870197 1.859409 1.820334 ... 2.304285 4.549563 -0.778416 1.500181 1.550724 1.195148 0.421326 4.889722 5.126897 4.747106
2018 2.404868 6.251701 2.815524 1.090624 1.348283 1.500539 0.678213 2.462246 2.322141 1.117817 ... 1.764937 5.496003 0.624486 0.453550 1.927713 0.772577 2.105525 2.875469 5.945421 4.076316

11 rows × 25 columns

In [21]:
usa = df.iloc[:11,1];china = df.iloc[:11,2];Russian = df.iloc[:11,3];Britain= df.iloc[:11,4];Australia=df.iloc[:11,5]
France=df.iloc[:11,6];Germany=df.iloc[:11,7];Korea= df.iloc[:11,8];Cuba=df.iloc[:11,9];
Italy,Japan,Ukraine,Canada,Spain,Brazil = df.iloc[:11,10],df.iloc[:11,11],df.iloc[:11,12],df.iloc[:11,13],df.iloc[:11,14],df.iloc[:11,15]
Netherlands,Hungary,Azerbaijan,Norway,Austria,Sweden= df.iloc[:11,16],df.iloc[:11,17],df.iloc[:11,18],df.iloc[:11,19],df.iloc[:11,20],df.iloc[:11,21]
Switzerland,Czech,Poland,Slovenia=df.iloc[:11,22],df.iloc[:11,23],df.iloc[:11,24],df.iloc[:11,25]
In [22]:
import plotly.graph_objs as go
import plotly.io as pio

years = [2008, 2009, 2010, 2011, 2012, 2013, 2014, 2015, 2016, 2017, 2018]

# define each line for each country
traces = [
    go.Scatter(x=years, y=usa, name='United States of America', mode='lines'),
    go.Scatter(x=years, y=china, name="People's Republic of China", mode='lines'),
    go.Scatter(x=years, y=Russian, name='Russian Federation', mode='lines'),
    go.Scatter(x=years, y=Britain, name='Britain', mode='lines'),
    go.Scatter(x=years, y=Australia, name='Australia', mode='lines'),
    go.Scatter(x=years, y=France, name='France', mode='lines'),
    go.Scatter(x=years, y=Germany, name='Germany', mode='lines'),
    go.Scatter(x=years, y=Korea, name='Korea', mode='lines'),
    go.Scatter(x=years, y=Cuba, name='Cuba', mode='lines'),
    go.Scatter(x=years, y=Italy, name='Italy', mode='lines'),
    go.Scatter(x=years, y=Japan, name='Japan', mode='lines'),
    go.Scatter(x=years, y=Ukraine, name='Ukraine', mode='lines'),
    go.Scatter(x=years, y=Canada, name='Canada', mode='lines'),
    go.Scatter(x=years, y=Spain, name='Spain', mode='lines'),
    go.Scatter(x=years, y=Brazil, name='Brazil', mode='lines'),
    go.Scatter(x=years, y=Netherlands, name='Netherlands', mode='lines'),
    go.Scatter(x=years, y=Hungary, name='Hungary', mode='lines'),
    go.Scatter(x=years, y=Azerbaijan, name='Azerbaijan', mode='lines'),
    go.Scatter(x=years, y=Norway, name='Norway', mode='lines'),
    go.Scatter(x=years, y=Austria, name='Austria', mode='lines'),
    go.Scatter(x=years, y=Sweden, name='Sweden', mode='lines'),
    go.Scatter(x=years, y=Switzerland, name='Switzerland', mode='lines'),
    go.Scatter(x=years, y=Czech, name='Czech', mode='lines'),
    go.Scatter(x=years, y=Poland, name='Poland', mode='lines'),
    go.Scatter(x=years, y=Slovenia, name='Slovenia', mode='lines'),
]

# layout
layout = go.Layout(
    title='GDP per capita growth rate of countries from 2008 to 2018',
    xaxis={'title': 'Year'},
    yaxis={'title': 'GDP per capita growth rate'},
    legend={'x': 1.02, 'y': 0.95},
    margin={'l': 50, 'b': 50, 't': 80, 'r': 50},
    hovermode='x'
)


fig = go.Figure(data=traces, layout=layout)


pio.show(fig)

                                          Figure 5

Resource: https://data.worldbank.org/indicator/NY.GDP.PCAP.KD.ZG

As can be seen from the line graph of the country's GDP per capita growth rate, the trends of GDP per capita growth rate tend to be the same in most countries during this decade (2008-2018), concentrated between -5 and 5. Among them, China's GDP per capita change is quite unique in that its separate line is shown above most of the other lines, remaining above 6 and even higher. Despite the low GDP per capita, China's high GDP per capita growth rate may explain his good performance in the Olympic Games. (Will continue analyzing this point later)

In [23]:
import numpy as np
import pandas as pd
df = pd.read_excel('323_data.xlsx', sheet_name='figure5')
# df.head(11)
y = df.iloc[0:89,1];
X = df.iloc[0:89,2:];
X = np.array(X)
y = np.array(y)
df.head(11)
Out[23]:
2008 MedalAmount GDPpercapita growthrate
0 USA 112 55271.656250 -0.820368
1 China 100 7412.874512 9.090280
2 Russian Federation 60 24887.853520 5.246210
3 Great Britain 51 43767.695310 -0.937078
4 Australia 46 44743.660160 1.513510
5 France 43 42996.210940 -0.302484
6 Germany 41 47643.222660 1.152030
7 Korea 32 32275.125000 2.233752
8 Italy 27 44623.601560 -1.615941
9 Japan 25 38781.199220 -1.272110
10 Ukraine 22 13719.271480 2.798642
In [24]:
# Linear regression 3D plot
import numpy as np
import plotly.graph_objs as go
from sklearn.linear_model import LinearRegression
# Fit the model
model = LinearRegression().fit(X, y)

print("Coefficients: ", model.coef_)
print("Intercept: ", model.intercept_)
print("Intercept: ", model.intercept_)
coef = model.coef_
intercept = model.intercept_
equation = 'y = {} + {}*x1 + {}*x2'.format(intercept, coef[0], coef[1])
print(equation)
# Generate prediction values
x_min, x_max = X[:, 0].min() - 1, X[:, 0].max() + 1
y_min, y_max = X[:, 1].min() - 1, X[:, 1].max() + 1
xx, yy = np.meshgrid(np.linspace(x_min, x_max, 10), np.linspace(y_min, y_max, 10))
Z = model.predict(np.c_[xx.ravel(), yy.ravel()])
Z = Z.reshape(xx.shape)

# Create the 3D plot
trace1 = go.Scatter3d(x=X[:, 0],y=X[:, 1],z=y,mode='markers',marker=dict(size=8,color=y,colorscale='Viridis',opacity=0.8))
trace2 = go.Surface(x=xx,y=yy,z=Z,colorscale='Viridis',opacity=0.6,)
data = [trace1, trace2]
layout = go.Layout(title={'text':'Multivariate linear fit plot of number of medals and GDP per capita and GDP growth rate','x':0.5,'y':0.97},
                   margin=dict(l=0, r=0, b=0, t=0),scene=dict(xaxis_title='GDP per caoita',yaxis_title='GDP growth rate',zaxis_title='Medal Amount'))
fig = go.Figure(data=data, layout=layout)
fig.show()

Coefficients:  [2.59868927e-05 9.87049373e-01]
Intercept:  26.803244458796424
Intercept:  26.803244458796424
y = 26.803244458796424 + 2.598689271355615e-05*x1 + 0.9870493725641499*x2

                                            Figure 6 (1)
In [25]:
import numpy as np
import plotly.graph_objs as go
from sklearn.pipeline import Pipeline
from sklearn.preprocessing import PolynomialFeatures
from sklearn.linear_model import LinearRegression

# Fit the model
model = Pipeline([('poly', PolynomialFeatures(degree=2)), ('linear', LinearRegression())])
model.fit(X, y)

# Print the equation of the fitted curve
coef = model.named_steps['linear'].coef_
intercept = model.named_steps['linear'].intercept_
equation = 'y = {:.2f} + {:.2f}*x1 + {:.2f}*x2 + {:.2f}*x1^2 + {:.2f}*x2^2 + {:.2f}x1x2'.format(intercept, coef[0], coef[1], coef[2], coef[3], coef[4])
print(equation)

# Generate prediction values
x_min, x_max = X[:, 0].min() - 1, X[:, 0].max() + 1
y_min, y_max = X[:, 1].min() - 1, X[:, 1].max() + 1
xx, yy = np.meshgrid(np.linspace(x_min, x_max, 10), np.linspace(y_min, y_max, 10))
Z = model.predict(np.c_[xx.ravel(), yy.ravel()])
Z = Z.reshape(xx.shape)

# Create the 3D plot
trace1 = go.Scatter3d(x=X[:, 0], y=X[:, 1], z=y, mode='markers',
                      marker=dict(size=8, color=y, colorscale='Viridis', opacity=0.8))
trace2 = go.Surface(x=xx, y=yy, z=Z, colorscale='Viridis', opacity=0.6,)
data = [trace1, trace2]
layout = go.Layout(title={'text':'Multivariate Nonlinear fit plot of number of medals and GDP per capita and GDP growth rate','x':0.5,'y':0.97},
                   margin=dict(l=0, r=0, b=0, t=0), scene=dict(xaxis_title='GDP per caoita',yaxis_title='GDP growth rate',zaxis_title='Medal Amount'))
fig = go.Figure(data=data, layout=layout)
fig.show()

y = 25.92 + 0.00*x1 + -0.00*x2 + 5.76*x1^2 + 0.00*x2^2 + -0.00x1x2

                                           Figure 6 (2)

We applied the linear regression and nonlinear regression, which are statistical modeling techniques used to examine the relationship between a dependent variable and independent variables (2 independent variables here). In the case of 3D plotting for linear regression, it can be used to create a regression plane that best fits the data points in a 3D scatter plot. It seems that as a country's GDP per capita and GDP per capita growth rate increase, so does its likelihood of winning Olympic medals. We’ve also checked nonlinear regression, which creates a regression surface that best fits the data points in a 3D scatter plot, even if the relationship is not linear. Nevertheless, the results are similar to linear regression, which shows the number of Olympic medals has a positive relationship with GDP per capita and GDP per capita growth rate. To sum up, from these two 3D scatter plots, we can suggests that economic development and growth are important factors that contribute to a country's success in the Olympic Games.

You might think that the positive relationship between GDP per capita/GDP per capita growth rate and the quantity of Olympic medals won is kind of intuitive and easy to speculate. Now, We want to dig a bit deeper: Why do some countries perform better in the Olympic Games (as measured by the number of medals)? What are the underlying reasons why countries with high GDP per capita usually have more medals? and How can we try to explain that correlation?

We will analyze these questions by exploring the following four different perspectives one by one. By doing so, we hope to provide a comprehensive understanding of the factors that contribute to a country's success in the Olympics, and how we can leverage this knowledge to improve athletic performance in countries with different levels of economic development.

1. Economic perspective¶

In [26]:
import pandas as pd
df = pd.read_excel('323_data.xlsx', sheet_name='Scale of Chinas sports industry',header =1 )
# y = df.iloc[0:89,1];
# X = df.iloc[0:89,2:];
# X = np.array(X)
# y = np.array(y)
df.head(3).set_index('Type')
Out[26]:
2015 2016 2017 2018 2019
Type
Industry Output (unit: 10^12 RMB) 1.71 1.90 2.20 2.66 2.95
Industry Output / GDP(%) 0.80 0.87 0.94 1.10 1.14
Industry Output per capita (RMB) 1240.00 1370.00 1546.00 1894.00 2094.00
In [27]:
import plotly.graph_objs as go

# data
years = [2015, 2016, 2017, 2018, 2019]
industry_output = [1.71, 1.9, 2.2, 2.66, 2.95]
industry_output_gdp = [0.8, 0.87, 0.94, 1.1, 1.14]
industry_output_per_capita = [1240, 1370, 1546, 1894, 2094]

industry_output_per_capita_k = [i/1000 for i in industry_output_per_capita]

trace1 = go.Scatter(x=years, y=industry_output, mode='lines+markers', name='Sport Industry Output (unit: 10^12 RMB)')
trace2 = go.Scatter(x=years, y=industry_output_gdp, mode='lines+markers', name='Sport Industry Output / GDP(%)')
trace3 = go.Scatter(x=years, y=industry_output_per_capita_k, mode='lines+markers', name='Sport Industry Output per capita (k RMB)')

# layout = go.Layout(title='Industry Output Trends',
#                    xaxis_title='Year',
#                    yaxis_title='Value',
#                    plot_bgcolor='rgb(255, 255, 255)')
layout = go.Layout(
    title={'text':'Sport Industry Output Trends of China','x':0.25, 'y':0.95},
    xaxis_title='Year',yaxis_title='Value',plot_bgcolor='rgb(240, 240, 240)',
    yaxis=dict(showgrid=True,gridcolor='rgba(0, 0, 0, 0.1)'
    )
)
fig = go.Figure(data=[trace1, trace2, trace3], layout=layout)
fig.show()

                                      Figure 7

Resource:

https://baijiahao.baidu.com/s?id=1760219976540280916&wfr=spider&for=pc

https://pdf.dfcfw.com/pdf/H3_AP202202171547603546_1.pdf?1645192405000.pdf

When we talk about how does economic perspective affect countries’ Olympic performance, we take China as a main discussing object. We can find an interesting phenomenon from the analysis above: despite having a lower GDP per capita than other countries, China still holds a higher percentage of medals. This raises the question of whether there is a bias in favor of China. To answer this question, One possible explanation is about China's GDP per capita growth rate, as shown in Figure 5, China's GDP per capita growth rate is consistently positive and higher than other countries’, while other countries' growth rate lines fluctuate. This suggests that GDP growth per capita may be a factor in China's success in winning medals.

Another possible worth mentioning reason why China has won quite lots of medals is that the difference in sports policies between China and Western countries. In Western countries, athletes typically have to pay for their own sports training fee, while in China, athletes are supported by the government: the government takes taxpayer money to train athletes. As China's GDP per capita increases, more money can be invested in sports, resulting in more resources being devoted to coaching, infrastructure, and other aspects of the sports industry. The figure 7 shows that the sports industry output in RMB, sports industry output over GDP percentage, and sports industry output per capita have all increased over time, indicating that more resources have been allocated to this particular industry each year, which makes China’s performance on Olympic games better.

2. Participant demographic perspective¶

In [28]:
import pandas as pd
df = pd.read_excel('323_data.xlsx', sheet_name='Athlets total number for summer',header=1)
df.head(11).set_index('Rank')
# y = df.iloc[0:89,1];
# X = df.iloc[0:89,2:];
# X = np.array(X)
# y = np.array(y) 

/opt/conda/lib/python3.10/site-packages/pandas/core/indexes/base.py:6982: FutureWarning:

In a future version, the Index constructor will not infer numeric dtypes when passed object-dtype sequences (matching Series behavior)
Out[28]:
Country Total Athlets Total medalist
Rank
1 United States of America 588 112
2 People's Republic of China 599 100
3 Russian Federation 454 60
4 Great Britain 309 51
5 Australia 432 46
6 France 309 43
7 Germany 420 41
8 Republic of Korea 265 32
9 Cuba 158 30
10 Italy 333 27
11 Japan 332 25
In [29]:
import plotly.graph_objs as go

# data
year_data = {
    2008: {
        'countries': ['United States of America', "People's Republic of China", 'Russian Federation','Great Britain', 'Australia', 'France', 'Germany', 'Republic of Korea', 'Cuba', 'Italy', 'Japan', 'Ukraine', 'Canada', 'Spain', 'Brazil']
,
        'athletes': df.iloc[0:15,2],'medalists': df.iloc[0:15,3]
    },
    2012: {
        'countries': ['United States of America', "People's Republic of China", 'Russian Federation', 'Great Britain', 'Germany', 'Japan', 'France', 'Australia', 'Republic of Korea', 'Italy', 'Netherlands', 'Ukraine', 'Hungary', 'Spain', 'Canada'],
        'athletes': df.iloc[19:34,2],'medalists': df.iloc[19:34,3]
    },
    2016: {
        'countries': ['United States of America', "People's Republic of China", 'Great Britain', 'Russia Federation', 'Germany', 'France', 'Japan', 'Australia', 'Italy', 'Canada', 'Republic of Korea', 'Netherlands', 'Brail', 'New Zealand', 'Azerbaijan'],
        'athletes': df.iloc[38:53,2],'medalists':df.iloc[38:53,3]
    }
}

def create_trace(year):
    trace = go.Scatter(
        x=year_data[year]['athletes'],y=year_data[year]['medalists'],mode='markers',
        marker=dict(size=15,color='#ADD8E6'), text=year_data[year]['countries'])
    return trace

fig = go.Figure()

for year in year_data.keys():
    trace = create_trace(year)
    fig.add_trace(trace)

updatemenu = []
buttons = []

for year in year_data.keys():
    button = dict(
        method='update',args=[{'visible': [year == y for y in year_data.keys()]}],label=str(year))
    buttons.append(button)

updatemenu = list([
    dict(active=0, buttons=buttons)
])

fig.update_layout(
    updatemenus=updatemenu,width=600, height=600,
    xaxis_title='Total Athletes',yaxis_title='Total Medalists',title='Olympic Medalists by Athletes, 2008-2016',
    plot_bgcolor='rgb(242, 242, 245)'
)

fig.show()

                                     Figure 8

Resource: https://www.nbcsports.com/northwest/tokyo-olympics/tokyo-olympics-numbers-participating-country-stats-and-facts

We are probably thinking that when the size of the athlete population in Olympics for a country increases, the higher probability this country will gain more medals. It is reasonable to assume that countries with higher GDP per capita would have more athletes participating in Olympics, as individuals with higher income tend to have more resources to invest in their favorite activities.

To investigate this percentage of Owning a medal, we analyzed data on the number of athletes and medals won in the three Summer Olympics using a scatter plot, which shown as figure 8. From the plot, we might revealed a positive linear relationship between the two variables across all three years of data, suggesting that countries with more athletes have a higher probability of winning medals.

3. Cultural perspective¶

Besides other perspectives, we also try to explore that whether sports cultures plays a crucial role in high-income countries through collecting Winter Olympic games data. We will discuss the following two main points using examples:

(1) Lifestyle aspect; for e.g. Europeans are very fond of participating in skiing activities in winter.¶

In [30]:
import pandas as pd
df = pd.read_excel('323_data.xlsx', sheet_name='figure9',header=0)
df.head(17).set_index('Number')

/opt/conda/lib/python3.10/site-packages/pandas/core/indexes/base.py:6982: FutureWarning:

In a future version, the Index constructor will not infer numeric dtypes when passed object-dtype sequences (matching Series behavior)
Out[30]:
Country Total number of ski resorts Share of people who ski 2010-2018total medal amount 2010 # of total mentals 2014 # of total mentals 2018 # of total mentals
Number
1 United States of America 525 88.0 37.0 28.0 23.0
2 Germany 680 0.18 80.0 30.0 19.0 31.0
3 Canada 297 80.0 26.0 25.0 29.0
4 Norway 172 0.22 88.0 23.0 26.0 39.0
5 Austria 442 0.34 47.0 16.0 17.0 14.0
6 Russian Federation 174 0.03 61.0 15.0 29.0 17.0
7 Republic of Korea 21 39.0 14.0 8.0 17.0
8 People's Republic of China 382 29.0 11.0 9.0 9.0
9 Sweden 256 0.18 40.0 11.0 15.0 14.0
10 France 249 0.13 41.0 11.0 15.0 15.0
11 Switzerland 358 0.35 35.0 9.0 11.0 15.0
12 Netherlands 61 52.0 8.0 24.0 20.0
13 Czech Republic 317 0.21 22.0 6.0 9.0 7.0
14 Poland 270 0.13 6.0 6.0 0.0 0.0
15 Italy 288 0.12 15.0 5.0 0.0 10.0
16 Slovenia 50 0.14 8.0 0.0 8.0 0.0
17 Spain 569 0.05 22.0 0.0 9.0 13.0

Scatter plot on the total number of medals in 3 winter Olympics versus share of people who ski for each country

In [31]:
import plotly.graph_objs as go


data = {
    "Country": ["Germany", "Norway", "Austria", "Russian Federation", "Sweden", "France", "Switzerland", "Czech Republic","Poland",
                "Italy", "Slovenia", "Spain" ],
    "Share of people who ski": [0.18, 0.22, 0.34, 0.03, 0.18, 0.13, 0.35, 0.21,0.13, 0.12, 0.14, 0.05],
    "2010-2018total medal amount": [80, 88, 47, 61, 40, 41, 35, 22, 6, 15, 8, 22]
}


fig = go.Figure(
    go.Scatter(
        x=data["Share of people who ski"],
        y=data["2010-2018total medal amount"],
        mode="markers",
        text=data["Country"],
        marker=dict(
            color="gray",
            size=15
        ),
        hovertemplate="Country: %{text}<br>" +
                      "Share of people who ski: %{x}<br>" +
                      "2010-2018total medal amount: %{y}"
    )
)


fig.update_layout(
    title={'text':"Scatter plot of Share of people who ski vs 2010-2018total medal amount",'x': 0.5,
        'y': 0.9,
        'xanchor': 'center',
        'yanchor': 'top'},
    xaxis_title="Share of people who ski",
    yaxis_title="2010-2018total medal amount",
)


fig.show()

                                               Figure 9

Resource: https://www.statista.com/statistics/801047/europe-share-of-population-skiing-by-country/

European ski participation rates, also named as ski population share, is calculated as the ski population divided by the total population, and we collected ski population share data from European countries. We hope to find that countries with ski population share may win more medals in the Winter Olympics.

Initially, we made scatter plot on the total number of medals in 3 Winter Olympics versus share of people who ski for each country. Unfortunately, there is no any clear relationship between these two variables. To explain this, we think it may be because some countries with a high share of people who ski do not participate in the Winter Olympics, and simultaneously, our ski population share data are mainly come from Europe, leading us for lacking data from non-European countries. In this way, many non-European countries that have been top 15 for three Winter Olympics have no relevant data (US, China, Canada, etc.), resulting in a somewhat fewer dots.

Bar chart for share of people who ski in European countries¶

In [33]:
import pandas as pd
df = pd.read_excel('323_data.xlsx', sheet_name='ski population share',header=0)
df.head(10).set_index('Number')

/opt/conda/lib/python3.10/site-packages/pandas/core/indexes/base.py:6982: FutureWarning:

In a future version, the Index constructor will not infer numeric dtypes when passed object-dtype sequences (matching Series behavior)
Out[33]:
European Countries Share of people who ski
Number
1 Liechitenstein 0.36
2 Switzerland 0.35
3 Austria 0.34
4 Finland 0.23
5 Norway 0.22
6 Czech Republic 0.21
7 Andorra 0.2
8 Slovakia 0.18
9 Sweden 0.18
10 Germany 0.18
In [34]:
import plotly.graph_objs as go


data = {
    "European Countries": df.iloc[0:43,1],
    "Share of people who ski": df.iloc[0:43,2]
}


fig = go.Figure(
    go.Bar(x=data["European Countries"],y=data["Share of people who ski"],
        marker=dict(
            color=['rgba(255, 215, 0, 0.8)' if country in ['Germany', 'Norway', 'Austria', 'Russian Federation', 'Sweden', 'France', 'Switzerland', 'Netherlands', 'Czech Republic', 'Poland', 'Italy', 'Slovenia', 'Spain', 'Serbia', 'Russia'] else 'rgba(173, 216, 230, 0.8)' for country in data["European Countries"]]
        )
    )
)


fig.update_layout(
    title={'text':"Bar chart of Share of people who ski in European Countries",'x': 0.5,
        'y': 0.9,'xanchor': 'center','yanchor': 'top'},
    xaxis_title="European Countries",
    yaxis_title="Share of People Who Ski",
)


fig.show()

                                             Figure 10

Nonetheless, in figure 10, the bar chart for share of people who ski in European countries, we can still see that the top 15 medal-winning countries (highlighted in yellow bars) in selected 3 Winter Olympics tend to be located in areas with higher ski share (ski participation rates). This further implies that the European countries that have been top 15 are also the countries that have a higher ski population share themselves.

(2). Cultural Strength; for e.g. China Table tennis¶

In [35]:
import pandas as pd
df = pd.read_excel('323_data.xlsx', sheet_name='China Table Tennis Competiton', header=0)
df.head(17)
Out[35]:
country Gold Silver Bronze China Winning Medal Proportion
0 2008 Summer Beijing - Table Tennis Medalist
1 Singles Men China China China 100%'
2 Singles Women China China China 100%'
3 Team Men China Germany Korea 1st
4 Team Women China Singapore Korea 1st
5
6 2012 Summer London - Table Tennis Medalist
7 Singles Men China China Germany 100%'
8 Singles Women China China Singapore 100%'
9 Team Men China Korea Germany 1st
10 Team Women China Japan Singapore 1st
11
12 2016 Summer Rio - Table Tennis Medalist
13 Singles Men China China Japan 100%'
14 Singles Women China China Korea 100%'
15 Team Men China Japan Germany 1st
16 Team Women China Germany Japan 1st

Resource: https://olympics.com/en/olympic-games

Some countries may have traditional sports that they are particularly good at and concentrate on. In these cuntries, with the presence of superior coaching resources, training infrastructure, this sport will naturally stand out. Surely we would expect that countries with higher GDP per capita are more likely to have the ability to invest more in this specific sport.

We can take China’s Table tennis as an example for this point. China has a long history of success in table tennis, and it is interesting to look at their ranking and medal count over the years. The table below summarizes China's medal count in three Summer Olympic games, showing that table tennis is definitely one of China's strengths.

It is worth noting that China only won two medals instead of three in the singles competitions at the 2012 London Olympics and 2016 Rio Olympics, meaning they did not take all of the three medals. The underlying reason is because after the 2008 Beijing Olympics, the International Table Tennis Federation decided to reduce the number of single table tennis players from each country from three to two. It simply means even though China won only two medals, its winning medals proportion was still 100%. Indeed, we have to say that this rule change limits the potential number of medals that China may win at their strength, but it also somewhat promotes other countries in table tennis.

Similar examples can be found in other sports, such as Russia's success in figure skating, which may be attributed to their cultural background and traditions in that sport.

4. Geographical perspective¶

Scatter plot on the total number of medals in 3 Winter Olympics versus the number of ski resorts for each country¶

In [36]:
import plotly.graph_objs as go
# Scatterplots on the total number of medals in 3 Winter Olympics versus the number of ski resorts for each country
df = pd.read_excel('323_data.xlsx', sheet_name='Number of Ski resort ', header=22)
df.iloc[:17,:4].set_index('number')
# df.iloc[:17,1]

/opt/conda/lib/python3.10/site-packages/pandas/core/indexes/base.py:6982: FutureWarning:

In a future version, the Index constructor will not infer numeric dtypes when passed object-dtype sequences (matching Series behavior)
Out[36]:
Country Total number of ski resorts 2010-2018 # of total mentals
number
1 United States of America 525 88.0
2 Germany 680 80.0
3 Canada 297 80.0
4 Norway 172 88.0
5 Austria 442 47.0
6 Russian Federation 174 61.0
7 Republic of Korea 21 39.0
8 People's Republic of China 382 29.0
9 Sweden 256 40.0
10 France 249 41.0
11 Switzerland 358 35.0
12 Netherlands 61 52.0
13 Czech Republic 317 22.0
14 Poland 270 6.0
15 Italy 288 15.0
16 Slovenia 50 8.0
17 Japan 569 22.0
In [37]:
import plotly.graph_objs as go
# Scatterplots on the total number of medals in 3 Winter Olympics versus the number of ski resorts for each country
df = pd.read_excel('323_data.xlsx', sheet_name='Number of Ski resort ', header=22)
data = {
#     "Country": ["United States of America", "Germany", "Canada", "Norway", "Austria", "Russian Federation", "Republic of Korea", "People's Republic of China", "Sweden", "France", "Switzerland", "Netherlands", "Czech Republic"],
#     "Total number of ski resorts": [525, 680, 297, 172, 442, 174, 21, 382, 256, 249, 358, 61, 317],
#     "2010-2018total medal amount": [88, 80, 80, 88, 47, 61, 39, 29, 40, 41, 35, 52, 22]
    "Country": df.iloc[:17,1],
    "Total number of ski resorts":df.iloc[:17,2],
    "2010-2018total medal amount":df.iloc[:17,3],
    
}

fig = go.Figure(
    go.Scatter(
        x=data["Total number of ski resorts"],
        y=data["2010-2018total medal amount"],
        mode="markers",
        marker=dict(color="lightblue",size=15),
        text=data["Country"],
        hovertemplate="Country: %{text}<br>" +
                      "Total number of ski resorts: %{x}<br>" +
                      "2010-2018total medal amount: %{y}"
    )
)

fig.update_layout(
    title={'text':"Scatter plot of Total number of ski resorts vs 2010-2018total medal amount",'x': 0.5,
        'y': 0.9,
        'xanchor': 'center',
        'yanchor': 'top'},
    xaxis_title="Total number of ski resorts",
    yaxis_title="2010-2018total medal amount",
)
fig.show()

                                                 Figure 11

Resource: https://www.skiresort.info/ski-resorts/

We expect that high-income countries have more resources to invest in sports facilities, to better use of potential geographical advantages, in contrast to low-income countries. To investigate our suspicions, we merged medal-winning countries with each country's total number of ski resorts. For this analysis, we collected data from the top 15 countries in three Winter Olympic Games (a total of 17 countries). We desire to find out if more ski resorts would make it easier for people to access ski sports, and thus be more likely to perform better in Winter Olympic sports.

From the scatter plot on the total number of medals in 3 Winter Olympics versus the number of ski resorts for each country, it seems that there is no clear relationship between the number of functioning ski resorts and the quantity of medals won. To explain this result, we think that other factors, such as the size and quality of ski resorts and participation rates, could also be relevant indicators. Nevertherless, our limited dataset prevents us from determining whether the number of ski resorts has any clear impact on a country's medal amounts.

In [38]:
import pandas as pd
df = pd.read_excel('323_data.xlsx', sheet_name='Number of Ski resort ', header=0)
df.head(17).set_index('number')

/opt/conda/lib/python3.10/site-packages/pandas/core/indexes/base.py:6982: FutureWarning:

In a future version, the Index constructor will not infer numeric dtypes when passed object-dtype sequences (matching Series behavior)
Out[38]:
Country Total number of ski resorts 2010 # of total mentals 2014 # of total mentals 2018 # of total mentals
number
1 United States of America 525 37 28.0 23.0
2 Germany 680 30 19.0 31.0
3 Canada 297 26 25.0 29.0
4 Norway 172 23 26.0 39.0
5 Austria 442 16 17.0 14.0
6 Russian Federation 174 15 29.0 17.0
7 Republic of Korea 21 14 8.0 17.0
8 People's Republic of China 382 11 9.0 9.0
9 Sweden 256 11 15.0 14.0
10 France 249 11 15.0 15.0
11 Switzerland 358 9 11.0 15.0
12 Netherlands 61 8 24.0 20.0
13 Czech Republic 317 6 9.0 7.0
14 Poland 270 6 NaN NaN
15 Italy 288 5 NaN 10.0
16 Slovenia 50 NaN 8.0 NaN
17 Japan 569 NaN 9.0 13.0
In [39]:
import plotly.express as px
df = pd.read_excel('323_data.xlsx', sheet_name='Number of Ski resort ', header=0)
data = {'Countries': df.iloc[:17,1].tolist(),
        'Number of Ski resort': df.iloc[:17,2].tolist()}
df = pd.DataFrame(data)

fig = px.choropleth(df, 
                    locations='Countries', 
                    locationmode='country names', 
                    color='Number of Ski resort', 
                    color_continuous_scale='YlGnBu',
                    range_color=[10, 700],
                    title='Number of Ski resort in European Countries',
                    width=1000,
                    height=800)
fig.show()

                                          Figure 12
In [40]:
import plotly.express as px
df = pd.read_excel('323_data.xlsx', sheet_name='Number of Ski resort ', header=22)
df.iloc[:17,:4].set_index('number')

/opt/conda/lib/python3.10/site-packages/pandas/core/indexes/base.py:6982: FutureWarning:

In a future version, the Index constructor will not infer numeric dtypes when passed object-dtype sequences (matching Series behavior)
Out[40]:
Country Total number of ski resorts 2010-2018 # of total mentals
number
1 United States of America 525 88.0
2 Germany 680 80.0
3 Canada 297 80.0
4 Norway 172 88.0
5 Austria 442 47.0
6 Russian Federation 174 61.0
7 Republic of Korea 21 39.0
8 People's Republic of China 382 29.0
9 Sweden 256 40.0
10 France 249 41.0
11 Switzerland 358 35.0
12 Netherlands 61 52.0
13 Czech Republic 317 22.0
14 Poland 270 6.0
15 Italy 288 15.0
16 Slovenia 50 8.0
17 Japan 569 22.0
In [41]:
import plotly.express as px
df = pd.read_excel('323_data.xlsx', sheet_name='Number of Ski resort ', header=22)
data = {'Countries': df.iloc[:17,1].tolist(),
        'The total number of medals': df.iloc[:17,3].tolist()}
df = pd.DataFrame(data)

fig = px.choropleth(df, 
                    locations='Countries', 
                    locationmode='country names', 
                    color='The total number of medals', 
                    color_continuous_scale='YlGnBu',
                    range_color=[5, 90],
                    title='The total number of medals won by countries ranking within the top 15 at the 2010 to 2018 Olympic Games.',
                    width=1000,
                    height=800)
fig.show()

                                             Figure 13

In these two thermodynamic diagrams, shown as figure 12 and figure 13, the number of ski resorts and medal-winning countries for top 15 in three Winter Olympic Games are displayed respectively. The shading of the diagrams indicates that darker colors represent countries with more ski resorts or more medals, while lighter colors represent countries with fewer ski resorts or fewer medals. Although the shades of the two thermodynamic diagrams do not correspond exactly, there is a clear correlation between the number of ski resorts and a country's performance in the Winter Olympics. The darker the color of an area showing the number of ski resorts, the darker the color of the corresponding area on the diagrams of the number of Winter Olympic medals. This suggests that geographical factors, such as the number of ski resorts, may play a role in a country's success in the Winter Olympics.

In addition, figure 13 also can be demonstrated that countries located at higher latitudes tend to have more medals, which shown as darker colors, suggesting better performance in the Winter Olympics. This pattern is linked to geographical factors as well; Intuitively, countries closer to the Arctic Circle typically experience colder temperatures and heavier snowfall, creating ideal conditions for winter sports and potentially enhancing athletes' ability to compete in the Winter Olympics. For example, several European countries, including Norway, Switzerland, and the Netherlands, have consistently ranked highly in the Winter Olympics medal list and are located at relatively high latitudes.

In [42]:
from IPython.display import Image

Image(filename='world climate zone.jpg', width=700)
Out[42]:
                                     Figure 14

Resource: https://qr.ae/prL3ts

From figure 14 shown above, we can see that with the change of latitude, the whole earth is divided into several different climate zones. We believe that this distribution of temperature zones is also an important geographic factor that can influence a country's sports performance in the Olympics, as well as its GDP per capita. In the following, we will explore this.

In [43]:
df = pd.read_excel('323_data.xlsx', sheet_name='Climate and latitute',header=1)
df.head(25).set_index('Country')
Out[43]:
Major Minor
Country
United States of America North Temperate Zone North Polar Zone
People's Republic of China North Temperate Zone North Tropical Zone
Russian Federation North Temperate Zone North Polar Zone
Great Britain North Temperate Zone Null
Australia South Temperate Zone Sorth Tropical Zone
France North Temperate Zone Null
Germany North Temperate Zone Null
Republic of Korea North Temperate Zone Null
Cuba North Tropical Zone Null
Italy North Temperate Zone Null
Japan North Temperate Zone Null
Ukraine North Temperate Zone Null
Canada North Temperate Zone North Polar Zone
Spain North Temperate Zone Null
Brazil South Tropical Zone South Temperate Zone
Netherlands North Temperate Zone Null
Hungary North Temperate Zone Null
Azerbaijan North Temperate Zone Null
Norway North Temperate Zone North Polar Zone
Austria North Temperate Zone Null
Sweden North Temperate Zone North Polar Zone
Switzerland North Temperate Zone Null
Czech Republic North Temperate Zone Null
Poland North Temperate Zone Null
Slovenia North Temperate Zone Null
In [44]:
import matplotlib.pyplot as plt

# data
countries = countries = ['United States of America', "People's Republic of China", 'Russian Federation','Great Britain',
                         'Australia', 'France', 'Germany', 'Republic of Korea',
                         'Cuba','Italy', 'Japan', 'Ukraine',
                         'Canada', 'Spain', 'Brazil', 'Netherlands',
                         'Hungary','Azerbaijan', 'Norway', 'Austria',
                         'Sweden', 'Switzerland', 'Czech Republic','Poland', 'Slovenia']
latitudes = ['North Temperate Zone', 'North Temperate Zone', 'North Temperate Zone', 'North Temperate Zone',
             'South Temperate Zone', 'North Temperate Zone', 'North Temperate Zone', 'North Temperate Zone',
             'North Tropical Zone', 'North Temperate Zone', 'North Temperate Zone', 'North Temperate Zone',
             'North Temperate Zone', 'North Temperate Zone', 'South Tropical Zone', 'North Temperate Zone',
             'North Temperate Zone', 'North Temperate Zone', 'North Temperate Zone', 'North Temperate Zone',
             'North Temperate Zone', 'North Temperate Zone', 'North Temperate Zone', 'North Temperate Zone',
             'North Temperate Zone']

# count the number of countries for each climate zone
low_lat_count1 = latitudes.count('North Tropical Zone')
low_lat_count2 = latitudes.count('South Tropical Zone')
low_lat_count = low_lat_count1+low_lat_count2

mid_lat_count1 = latitudes.count('North Temperate Zone')
mid_lat_count2 = latitudes.count('South Temperate Zone')
mid_lat_count = mid_lat_count1 +mid_lat_count2

high_lat_count = len(countries) - low_lat_count - mid_lat_count
# print(low_lat_count,mid_lat_count,high_lat_count)

# create bar chart
fig, (ax1, ax2) = plt.subplots(1, 2, figsize=(13, 6))

x = ['Tropical', 'Temperate', 'Polar']
y = [low_lat_count, mid_lat_count, high_lat_count]
colors = ['red', '#98FB98', '#87CEEB']
ax1.bar(x, y, color=colors)

ax1.set_title('Bar Chart of Countries\' Latitude Distribution', fontsize=17)
ax1.set_xlabel('Major Zone',fontsize=14)
ax1.set_ylabel('Number of Countries')

# create pie chart
labels = ['Tropical', 'Temperate', 'Polar']
sizes = [low_lat_count, mid_lat_count, high_lat_count]
colors = ['red', '#98FB98', '#87CEEB']
explode = (0.1, 0, 0)
ax2.pie(sizes, explode=explode, labels=labels, colors=colors,
        autopct='%1.1f%%', startangle=90)

ax2.set_title('Pie chart of Countries\' Latitude Distribution', fontsize=17)

plt.show()

                                           Figure 15

Resource: Wikipedia for each country, and you are welcome to explore it yourself and find out we're right! :)

What insights can we glean if we explore the latitudes of all top 15 ranked countries in the all past six Olympic game? In figure 15, we have collected information on the climate zones of the top 15 countries in the last six Olympic Games: Tropical zone, North Temperate zone, South Temperate zone, North polar zone and South polar zone. (We do not distinguish between the northern and southern hemispheres in the graphs) We have chosen the predominant climate area within each country as a major consideration, and exclude minor climate zones since smaller portions of a country's land area tend to hold less significance. In this way, we can draw more a plausible estimate. In the pie chart, the result reveals that most of the top 15 ranked countries in the past six Olympic Games are located in the (North) temperate zone, and same result as shown in bar chart. This may indicate that areas with milder climates are more suitable for sports, at the same time the data shown earlier (figure 1, figure 13, etc) also shows that the economic development of the temperate zone is relatively better, with a higher GDP per capita.

Limitations and Further Research:¶

  1. Lack of precise data on the number of ski resorts and ski population share in countries that ranked in the top 15 in the selected three Winter Olympics, which may lead to inaccuracies in the scatter plots we made.
  • To be specific, limited ski population share data, as it only comes from European countries, resulting in a lack of data from non-European countries, such as the US, China, and Canada. etc. This causes results in somewhat fewer dots shown. In this way, we might be unable to see the clear underlying relationship between the number of medals in the Winter Olympics versus the share of people who ski. Moreover, for the scatter plot on the total number of medals in the winter Olympics versus the number of ski resorts, because we did not account for other potentially relevant factors in the dataset, such as the size and quality of ski resorts and participation rates, the possible potential relationship is then not shown.

  1. Our 1st selected Olympics data started in 2008, however, for sports industry output data, we only find data from 2015 and ended in 2019 in China. Due to the lack of time series data on sports investment since 2008 and the absence of year-by-year data for other countries, we cannot determine whether the trend of sports investment in the 2015-2019 period is similar to the 2008-2018 period in China, nor can we be sure about the development of the sports sector in other countries.

  2. To explore the relationship between countries' Olympic performance (measured by medal counts) and their GDP per capita, a viable approach that we did not include here would be to identify a policy experiment that would generate variation in countries' GDPs and examine its impact on Olympic performance. Unfortunately, due to the project’s time constraints, we were unable to identify such a policy experiment and investigate its effectiveness.

Considering the limitations of our current study, future research should focus on collecting more precise data on ski resorts and ski population shares worldwide, including non-European countries, to better understand the relationship between winter sports participation and Olympic success. Additionally, acquiring time series data on sports investment since 2008 for various countries would enable a more accurate comparison of sports development investments and their potential impact on Olympic performance. Lastly, analyzing a policy experiment that affects countries' GDPs could provide insights into the causal relationship between GDP per capita and Olympic success, further enhancing our understanding of the factors influencing Olympic performance.

Conclusion:¶

To investigate the relationship between a country's GDP per capita and the number of Olympic medals won, we collected data related to the Olympic Games before the COVID-19 pandemic. Our analysis focused on the medals earned in three summers and three winter Olympic Games.

Our findings indicate that there is a positive correlation between a country's GDP per capita and the number of medals won in both linear and non-linear regression models. Additionally, the GDP per capita growth rate is also positively associated with the country’s Olympic medals amounts in the Olympic Games. We further explored various aspects might contribute to higher-income countries achieving higher rankings in the Olympic Games.

China's consistent progress in the Summer Olympics can be attributed to its high GDP per capita growth rate. As GDP increases, more funding is allocated to sports, and China's policy emphasizes government support for nurturing athletes. Industry output data shows that China's sports output has been growing annually. High-income countries tend to have more athletes and win more medals due to a larger participation base population. In low-income countries, people may prioritize work over being a athletes. From the scatter plot data, there is a linear relationship between the participation base population and the number of medals won. As a result, high-income countries tend to have relatively more athletes, which increases their likelihood of winning medals.

We observed that the winners of the Winter and Summer Olympics tended to fall into two distinct country groups. With the winning countries of the Winter Olympics primarily concentrated in Europe, we wondered if certain sports might be related to national geographic characteristics. High-income countries are typically more willing to invest in sports, to better use potential geographical advantages, and construct more accessible sports places.

We also examined the number of ski resorts and ski population share in the top 15 winning countries across three Winter Olympics. However, our analysis revealed no clear relationship between these 2 variables and the number of medals won, which may be due to our limited dataset does not allow us to establish a definitive connection directly.

In [ ]: